I had a chance to analyse the WeRateDogs, a Twitter account that rates dogs. The ratings have a denominator of 10 but the numerator changes depending on how good the dog is,something like 10/10 or 14/10.
My objective was to get the dog with the highest rating, get to know the most popular dog through the retweets and favorite counts in the twitter handles.
And pheeew!! After a serious analysis of the WeRateDogs twitter posts,our top two rated beauties are here:
Duddles
The favourites of all. The most with favorite counts. And they are of the English_springer dog breeds.
Still trying to believe he's one of the lowest-rated. He drives!
There is strong positive correlation between Favorites counts and retweets counts. I would easily use either for dog popularity.
The chart bellow shows the positive correlation.
However, a relationship between either the retweet counts or favorite counts proved futile. They are not correlated in any way. A dog's rating could not be based on favorite nor retweet counts.
An example of a correlation scatter plot:
This was such an amazing project,with indepth analysis and inferensing on the datasets from the WeRatedogs posts on twitter.